2011 ACGME Residency Work Restrictions Associated with Increased Inpatient Mortality at Teaching Institutions

The Accreditation Council for Graduate Medical Education (ACGME) mandated sweeping restrictions to resident physician work hours in 2003. Hospitals saw increased costs for night coverage and longer hours for attending physicians, with unclear results for patient safety. On July 1, 2011, the ACGME required hospitals to increase rest periods between duty hours and no more than 16 hours of continuous intern physician duty.

We utilized the Nationwide Inpatient Sample (NIS), a nationally representative 20% stratified sample of all-payer inpatient discharges for nonfederal hospitals, comparing inpatient mortality in the second half of 2011, relative to the first six months in teaching and non-teaching hospitals. Patient age, gender, race, comorbidities, Medicaid status, hospital discharge volume, hospital teaching status, and urban vs. rural location were independent variables based on Anderson-Newman criteria for health care utilization. Patients over 110 years of age were excluded as possible coding errors. A difference-in-difference model was evaluated using multiple logistic regression with inpatient death as the dependent variable and the effect of residency restrictions determined as an interaction term between teaching hospital status and discharge on or after July 1. Adjusted and unadjusted marginal effects were calculated. The same analysis was performed on 2010 data.

Compared with non-teaching hospitals, teaching hospitals had younger patients, much higher annual discharge volumes, and were more often in an urban setting (descriptive statistics not shown). Absolute inpatient mortality declined for both teaching and non-teaching hospitals in the second half of the year for both 2010 and 2011. The marginal decline in mortality for teaching hospitals after July 1, 2011 was 0.1% less than non-teaching hospitals, giving a relative change of 5.34%, adjusted OR 1.05 (p=0.001). This effect was not seen in 2010.

We show a significant effect of teaching hospitals in 2011 adversely impacting inpatient mortality after July 1, that is absent during the same time frame in 2010. This finding is robust to adjustment for demographic, clinical, and hospital characteristics.

This study highlights the effect of the most recent ACGME work hour restrictions on patient safety, suggesting a tradeoff between fatigue and continuity of care. In a cross-over study of medical residents implementing the 2011 ACGME rules, the mean number of handoffs increased from 3 to 9 over 72 hours, increasing opportunities for errors. If 98% of information is maintained between each handoff, only 83% of the original information is preserved after nine handoffs, compared to 94% after three.

 This study has limitations. The effect may be related to some other factor coincident with the ACGME changes. Some patients discharged in the second half of 2011 would have been treated in the first half, and new work hour restrictions may have been implemented early. Comparing 2010 to 2012, dropping 2011, would remove the “transition” effect. Future work includes validating these findings using the AMA-FREIDA residency database to analyze impact specific to relevant condition and specialty.

If these trends continue, the ACGME should revisit the 2011 work hour requirements and implement changes to improve patient safety in hospitals.